arXiv cs.AI· Songping Wang, Peiyang Li, Yi Huang, Fengting Li, Xuhao Ran, Yueming Lyu, Chao Feng, Caifeng Shan·· 12 小时前AI 评分44
MobileInject：评估跨模态一致性注入下 MLLM 移动智能体的推理到行动安全
MobileInject: Evaluating Reasoning-to-Action Security of MLLM-Based Mobile Agents under Cross-Modal Consistency Injection
AI 导读
研究者提出 MobileInject，首个评估 MLLM 移动智能体多模态注入攻击与推理到行动安全风险的基准，覆盖 11 个场景、9 类安全风险和 108 个可执行风险实例。该基准引入跨模态一致性注入，并配套多角色协同欺骗与推理伪装注入两种增强攻击方法，以及推理安全诊断框架。系统实验显示移动智能体普遍易受多模态注入攻击，许多失败源于早期阶段的推理劫持。
正文 · 原文
Abstract:Multimodal large language model (MLLM)-based mobile agents are rapidly evolving from conversational assistants into autonomous systems that perceive, reason over, and act on smartphone environments. This capability also exposes a new security risk: adversaries can embed malicious objectives in cross-application mobile content and steer agent behavior through indirect multimodal injection. Although prompt injection has been studied in desktop and web settings, its cross-modal properties in mobile ecosystems remain poorly understood, and existing security analyses are still limited. To close this gap, we introduce MobileInject, the first benchmark for evaluating mobile agents against multimodal injection attacks and reasoning-to-action security risks. MobileInject provides a high-fidelity interactive mobile evaluation framework covering 11 scenarios, 9 security-risk categories, and 108 executable risk instances. It further introduces Cross-Modal Consistency Injection to exploit corroboration across modalities, two enhanced attack methods, Multi-Role Collaborative Deception and Reasoning-Camouflage Injection, to intensify trust manipulation and reasoning hijacking, and a Reasoning Security Diagnosis Framework for fine-grained failure analysis. Systematic experiments show that mobile agents are broadly vulnerable to multimodal injection, and that many failures stem from reasoning hijacking at early stages. These findings establish MobileInject as a systematic cross-modal benchmark for studying and improving the security of mobile agents. Code release details are provided in the Appendix Open Science section.
| Subjects: | Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.23307 [cs.CR] |
| (or arXiv:2610.23307v1 [cs.CR] for this version) | |
| https://doi.org/10.48550/arXiv.2610.23307 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Songping Wang [view email]
[v1]
Fri, 16 Oct 2026 11:10:30 UTC (5,933 KB)
来源：arXiv cs.AI · arxiv.org